Tolerance Stack-Up Analysis

Tolerance stack-up analysis examines how individual component variations combine in an assembly. Every manufactured part has dimensional variation within its specified tolerance. When multiple parts fit together, their tolerances accumulate — potentially causing the assembly to fail even though every individual part is 'within spec.' If five parts each vary by ±0.5mm, the worst-case stack-up is ±2.5mm, which may exceed the assembly's functional requirement. Engineers use two approaches: worst-case analysis (assumes all parts are simultaneously at their extreme limits — very conservative) and statistical analysis (assumes variations are normally distributed and unlikely to all be at extremes simultaneously — more realistic). The practitioner's insight: tolerance stack-up is often where designs fail in production.

When to use it

When designing assemblies or processes with multiple interacting components, estimating total project uncertainty from individual task uncertainties, analyzing quality problems that emerge in production but not prototyping, or understanding why systems with many variable components produce more extreme outcomes than expected.

How it can help

Tolerance stack-up thinking applies to any process where multiple sources of variation combine. In project management, each task's time estimate has uncertainty — the total project duration isn't the sum of 'expected' task durations but the stack-up of all uncertainties. This is why projects consistently overrun: managers add expected task times without accounting for variance accumulation. In supply chains, each supplier's delivery variation compounds through the chain. In communication, each retelling introduces variation — after multiple handoffs, the message may be unrecognizable.